Research map / Detection and segmentation / CNN features
Deep Grouping Model for Unified Perceptual Parsing
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add DGM to ResNet50 | Unified parsing model · Broden+ O+P+S | Object mIoU: 23.83 -> 24.76 | Helped |
| Add DGM to ResNet50 | Unified parsing model · Broden+ O+P+S | Object P.A.: 77.23 -> 75.15 | Hurt |
| Add DGM to ResNet50 | Unified parsing model · Broden+ O+P+S | Part mIoU: 30.10 -> 31.26 | Helped |
| Add DGM to ResNet50 | Unified parsing model · Broden+ O+P+S+M+T | Texture Top-1 Acc.: 57.44$^*$ -> 60.10 | Helped |
| Add DGM to UPerNet | UPerNet · ADE20k validation | mIoU: 42.66 -> 43.51 | Helped |
| Add DGM to HRNetv2 | HRNetv2 · ADE20k validation | mIoU: 43.20 -> 43.46 | Helped |
| Add DGM to DeepLabV3 | DeepLabV3 · ADE20k validation | mIoU: 44.1 -> 44.86 | Helped |
| Add TDMP (ablation) | DGM · Broden+ O+P+S | Object mIoU: 24.58 -> 24.76 | Helped |
| Add TDMP (ablation) | DGM · Broden+ O+P+S+M+T | Part mIoU: 29.94 -> 30.28 | Helped |
| Add TDMP (ablation) | DGM with UPerNet backbone · ADE20k validation | mIoU: 43.64 -> 43.51 | Hurt |
| Add TDMP (ablation) | DGM with HRNetv2 backbone · ADE20k validation | mIoU: 43.86 -> 43.46 | Hurt |
| Add TDMP (ablation) | DGM with DeepLabV3 backbone · ADE20k validation | mIoU: 44.31 -> 44.86 | Helped |
Closest papers by meaning
- Perceptual Group Tokenizer: Building Perception with Iterative Grouping (ICLR 2024 · similarity 0.89)
- Latent Noise Segmentation: How Neural Noise Leads to the Emergence of Segmentation and Grouping (ICML 2024 · similarity 0.86)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · similarity 0.86)
- Learning Hierarchical Image Segmentation For Recognition and By Recognition (ICLR 2024 · similarity 0.85)
- Calibrating Multi-modal Representations: A Pursuit of Group Robustness without Annotations (CVPR 2024 · similarity 0.85)
- Native Segmentation Vision Transformers (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.